Application of genetic algorithm to job shop scheduling problems with active schedule constructive crossover

This paper explains the application of a genetic algorithm (GA) to job shop scheduling problems. Given combinatorial problems, it is important to develop an efficient representational scheme and effective genetic operators of the GA for better performance. Many representational schemes exist to encode ordering information of jobs and/or operations, which have their own advantages and shortcomings. For better performance, we use both job-ordered list on each machine and operation list representational schemes for a crossover and mutations, respectively, and develop a new crossover, so called, an active schedule constructive crossover (ASCX) and two mutations. This crossover can exchange meaningful ordering information of parents effectively without producing illegal solutions and the mutations can easily provide various operation orderings. Simulation results on five benchmark problems show that our genetic operators are very powerful and suitable to job shop scheduling problems and our GA outperforms the previous GA-based approaches and the GA with position-based crossover (PBX)